Abstract
A retrospective study was performed to characterize 64 cases of anal sac gland carcinoma (ASGC) in cats. All ASGCs diagnosed between 1995 and 2007 at a private diagnostic laboratory in the UK were reviewed. Apocrine gland origin was confirmed in a subset of these tumors by immunohistochemistry and the use of the glandular cytokeratin antibody (CAM 5.2). Associated clinical, gross, and histologic features were compared with those of canine ASGC. Anal sac gland carcinoma accounted for 0.5% of all feline skin neoplasms. Thirty-nine of the cats with ASGC were female, with a female: male ratio of 1.56. Fifty-two (81.1%) of the 64 tumors were in Domestic Shorthair cats, 5 (7.8%) in Siamese, 3 (4.8%) in Domestic Longhair, 2 (3.1%) in Burmese, and 1 (1.6%) each in a Birman and a Persian cat. Significant differences in prevalence of ASGC among breeds were not detected. Cats ranged in age from 6 to 17 years (median and mean age, 12 years). More than three quarters of the affected cats for which postsurgical outcome was known were euthanatized or died as a direct consequence of the neoplasm, with a median survival of 3 months. Survival rates at 1 and 2 years were 19 and 0%, respectively.
The anal sacs are paired cutaneous anal diverticula that occur within the ventrolateral perianal region of carnivores and many rodents. 2,5 The anal sacs are lined by a cornifying stratified squamous epithelium and located between the smooth muscle internal anal sphincter and skeletal muscle external anal sphincter. The excretory duct of each sac opens onto the keratinized anal junction. In the cat, the anal sacs are surrounded by tubular apocrine glands and sebaceous glands, whereas only apocrine glands are found in dogs. 2
Anal sac gland carcinoma (ASGC), also known as adenocarcinoma of the apocrine glands of the anal sac, is a malignant tumor arising within the apocrine secretory epithelium associated with the wall of the anal sac. Although uncommon, this neoplasm is well recognized in the dog, representing the most frequent malignant tumor of the perineum (approximately 2% of canine skin tumors), with a peak incidence between 7 and 12 years of age. 8,9,26 Some earlier reports have indicated a female predisposition, whereas others have shown no apparent gender disparity. 21,25
Tumors are usually unilateral and present as poorly circumscribed swellings or infiltrative masses; however, smaller circumscribed nodules can mimic the gross appearance of hepatoid gland adenomas. Most tumors are detected via digital rectal examination, but this neoplasm can present as an occult disease, necessitating careful rectal palpation. Tumors exhibit variable growth but tend to be highly malignant, with invasion into the deeper perirectal tissues, invoking a desmoplastic (scirrhous) response and early metastasis via lymphatic drainage to the regional sacroiliac and sublumbar lymph nodes. Subsequent dissemination to internal organs (e.g., lungs, liver, and spleen) occurs less frequently. Large tumors can impinge upon the rectum, resulting in tenesmus and dyschezia, but ulceration is an uncommon complication in dogs. Surgical excision can prove curative for some of the smaller canine lesions; however, neoplasms with documented lymphatic metastases are associated with a poor prognosis. 8,9,26 A significant proportion of animals develop systemic clinical signs (polyuria, polydypsia, weakness, vomiting, and anorexia) attributable to humoral hypercalcemia of malignancy from secretion of parathyroid hormone-related protein (PTHrP) by the neoplastic cells. 6,13,16
Despite considerable knowledge of canine ASGC, this tumor rarely has been documented in cats; to the authors' knowledge, only 2 cases have been reported. 15,18 Thus, little is known of the clinical features or biological behavior of feline ASGC. The purpose of this study was to characterize the signalment, clinical, gross, and histopathologic characteristics of anal sac gland carcinoma in cats, to compare these features with canine ASGC, and to investigate the diagnostic utility of an immunohistochemical glandular epithelial cytokeratin marker (CAM 5.2).
Materials and Methods
Animals
Formalin-fixed tissues from the 64 cats were received by Finn Pathologists (Weybread, Diss, Norfolk, UK) for histopathologic examination from 1 January 1995 to 31 December 2007. Tissues were generally submitted as surgical biopsies from cats that had a history of a tumor involving the anal/perineal region. In all cases, the examining pathologist had made either a definitive or, rarely, presumptive diagnosis of ASGC. Archived embedded tissues were retrieved, and freshly stained slides were prepared and evaluated by a board-certified pathologist (A. M. Shoieb) to confirm the diagnosis.
Follow-up information for each case (either included with the original submission form or collected from submitting veterinary practices via telephone) included: signalment (breed, gender, neutering, age at tumor diagnosis); size, location, and appearance of the tumor; clinical signs associated with the neoplasm; type of surgery (incisional or excisional biopsy); ancillary treatment; and clinical outcome, including postoperative complications, disease-free interval, and, if applicable, the cause of death.
Database records were subsequently analyzed to derive demographic information regarding the overall prevalence of ASGC in the UK feline population and to assist in identifying possible breed and gender predilections or any bias in the geographic distribution of cases. Statistical evaluations of tumor prevalence in relation to breeds and gender included in this study were conducted by chi-square tests in the NCSS 2007 software package (http://www.ncss.com). Differences were considered significant at P ≤ .05.
Histology
All tissue samples were fixed in 10% neutral buffered formalin. Selected tissues were embedded in paraffin, sectioned at 5 µm, and mounted on positive-charged glass slides (Superfrost/plus, Fischer Scientific). Tissue sections were stained with HE.
Immunohistochemistry
Immunohistochemistry was performed with the EnVision/horseradish peroxidase system (Dako), with anti-cytokeratin monoclonal antibody CAM 5.2. 1,10,14,17,19 CAM 5.2 binds specifically to cytokeratin peptide 8 in tissue sections. 10,19 Sections were deparaffinized by immersion in 3 changes of xylene and rehydrated in descending alcohol solutions. Slides were washed in tap water. Endogenous peroxidase activity was quenched by incubating the slides in 2.0% hydrogen peroxide for 20 minutes at room temperature, then rinsing the slides in tap water, then tris-buffered saline (TBS), pH 7.6. The sections were digested with 0.01% Trypsin II in distilled water for 10 minutes at room temperature (Becton Dickinson) then washed in tap water for 5 minutes, followed by TBS for 5 minutes. Nonspecific binding was blocked with normal goat serum (blocking serum), 1 : 10, for 20 minutes in a moist chamber. Slides were then incubated for 40 minutes with diluted mouse monoclonal antibody CAM 5.2 (Becton Dickinson) or negative control reagent. The slides were washed in TBS for 5 minutes then incubated with goat anti-mouse immunoglobulin (Dako EnVision) in a moist chamber at room temperature for 45 minutes, rinsed with TBS for 10 minutes, and incubated for 10 minutes with diaminobenzidine chromogen (Dako) at room temperature. The color change was monitored on positive control slides and stopped by immersing all slides in deionized water. Slides were counterstained by incubating with Mayer's hematoxylin for 40 seconds then washed with running tap water for 5 minutes. Slides were dehydrated through ascending alcohol solutions, cleaned in xylene, and coverslipped.
Appropriate positive controls for the antibody were run concurrently and showed appropriate immunoreactivity. Negative control sections for immunohistochemistry were processed without the primary antibody.
Results
Signalment
Fifty-two of 64 cats (81.1%) were in Domestic Shorthair (DSH) cats, 5 of 64 (7.8%) in Siamese, 3 of 64 (4.8%) in Domestic Longhair (DLH), 2 of 64 (3.1%) in Burmese, and 1 (1.6%) each in a Birman and a Persian cat. Thirty-nine cats (61%) with ASGC were female; of these, 34 (87.2%) had been spayed. Twenty-five (39%) were male, and 23 of 25 (92%) had been castrated. The cats ranged in age from 6 to 17 years; median and mean ages were both 12 years.
Demographics
The 64 cases of ASGC were derived from 121,640 feline biopsy accessions received during the 13-year study period, of which approximately 11% (13,023) were recorded as skin neoplasms. The 64 cases accounted for 0.5% of feline skin neoplasms and 0.05% of all feline accessions (Table 1). Sixty of the cases were submitted by veterinary practices within England; 2 cases each were from Scotland and Ireland. Subjectively, the distribution of cases appeared to correlate with the geographic spread of veterinary practices serviced by the laboratory. Two practices in the counties of Kent and the West Midlands submitted 2 cases each; the remaining 60 cases were submitted from 60 different practices.
Breed and gender of cats with anal sac gland carcinoma, combined skin neoplasms, and total biopsy accessions.∗
∗ASGC = anal sac gland carcinoma; DSH = Domestic Shorthair cat; DLH = Domestic Longhair cat.
The gender ratio (female : male) for all feline skin neoplasms (6,382 : 6,641 = 0.96) and for total feline accessions (60,481 : 61,159 = 0.99) was approximately 1 (Table 1). In contrast, the female : male ratio for ASGC was 39 : 25 = 1.56. Analysis of breed distribution of cats with ASGC versus cats with skin neoplasms and total feline biopsy accessions (Table 1) revealed the following breeds to be overrepresented: Siamese, Burmese, Birman, and Persian. The DLH was underrepresented. The proportion of DSH cats among cats with ASGC approximated that for all feline skin neoplasms and total feline accessions. The gender and breed distribution of ASGC cases did not differ significantly from the gender and breed distribution of the reference feline population.
Clinical signs
The most common clinical signs in cats with ASGC included dyschezia, recurrent constipation, change in volume or character of feces, and perineal swelling, ulceration, or both, sometimes accompanied by purulent to hemorrhagic discharge. At least 7 cats were presented for veterinary attention after sudden discovery of a perianal lesion by the owner, with either abrupt onset of clinical signs (perianal discomfort, abnormal tail carriage, blood around the perianal region) or without observation of clinical signs. In one cat, the tumor was discovered during routine veterinary examination.
Attending veterinarians generally reported a firm nodular perianal mass, ulceration with a fistulous tract, or both. Most tumors appeared to be associated with the anal sac and were often presumptively diagnosed as an anal sac abscess. Response to antibiotic treatment was minimal to none, however. Two cases were reported as rectal and vulvar in origin, respectively. Enlargement of sublumbar lymph nodes was documented radiographically in 2 cases.
Serum calcium concentration was measured in only 5 cases; in 4 cats, concentration was within the reference range; 1 cat had a marginal hypercalcemia (2.6 mmol/L; laboratory reference range, 1.6–2.5 mmol/L) but reportedly exhibited no clinical signs attributable to hypercalcemia.
Two cats had mild to moderate azotemia; 1 cat had mild hyperglobulinemia; concurrent hyperthyroidism was recorded in 2 cats; and uremia was documented in 1 cat with chronic renal failure. Clinical signs referable to sciatic paralysis were detected in 1 cat.
Gross
The primary tumor most commonly appeared as a deep-seated, firm, nodular perianal mass ventrolateral to the anus and involving, or within close proximity to, the anal sac. Tumors were unilateral and typically 1–2 cm in diameter (range, 0.5–5 cm). Location was reported as left in 21 cats, right in 15, and unrecorded or described as ventral in 28.
Eighteen tumors (28%) were associated with a purulent to hemorrhagic and purulent anal sac discharge and communicating fistula; ulceration was a common complication. Tumors were white to tan, sometimes hemorrhagic, and frequently lobulated, with occasional small cysts on the cut surface. Perianal tissue surrounding the neoplasm was often thickened and firm. There was occasional invasion into the wall of the rectum or anus, but luminal compression was not readily recognized.
Histopathology
Microscopically, the tumors had typical carcinomatous features, with solid to glandular components. In all cases, the subepithelial connective tissue and normal apocrine glands of the anal sac wall were replaced by an unencapsulated, multilobulated, infiltrative, densely cellular neoplasm (Fig. 1).
Anal sac gland carcinoma; cat. An expansile and invasive tumor (T) arising adjacent to the anal sac wall (arrows = epithelium; asterisk = lumen) compresses and displaces skeletal muscle of the external anal sphincter. HE stain. Bar = 0.3 mm.
Tumor lobules were composed of epithelial cells arranged in nests, packets, cords, and (in more differentiated areas) tubuloacinar structures, separated and supported by a fine fibrovascular stroma. Neoplastic cells sometimes formed rosettes. The cells were polygonal to columnar with variably distinct cell borders and a moderate amount of eosinophilic granular cytoplasm. Nuclei were round to oval and basilar (within tubuloacinar structures) with mild anisokaryosis, coarsely stippled chromatin, and 1 or 2 nucleoli. The mitotic index was moderate, with 1–4 mitotic figures in many high-power fields. There was sparse single cell necrosis, occasional amorphous eosinophilic material (secretory product) in tubuloacinar lumina (Fig. 2), and low to moderate numbers of lymphocytes, plasma cells, and hemosiderin-laden macrophages, with occasional small numbers of eosinophils, and neutrophils, scattered within the connective tissue stroma. Similar inflammatory cells infiltrated the subepithelial connective tissue of the anal sac wall, and remaining apocrine glands were moderately ectatic, hyperplastic, or both.
Anal sac gland carcinoma; cat. Acinar pattern: acini are lined by one or more layers of cuboidal epithelial cells and contain small amounts of granular to amorphous eosinophilic secretory product (asterisks). HE stain. Bar = 20 µm.
Nests of tumor cells and individual tumor cells infiltrated the fibrovascular stroma and surrounding perirectal skeletal muscle of the external anal sphincter with scirrhous response. Lymphatic invasion was evident in tissue sections from 7 cats; invasion of the excised sublumbar lymph node was detected in 1 cat.
Immunohistochemistry
A randomly selected subset of tumors (n = 12) was reacted with CAM 5.2. Consistent positive cytoplasmic reactivity of mild to marked intensity was evident within neoplastic epithelial cells (Fig. 3). Nonneoplastic apocrine glands within the anal sac wall also were immunoreactive (Fig. 4), whereas the stratified squamous epithelium of the anal sac lining, epidermis, follicular epithelium, and cutaneous sebaceous glands, when present in the section, were negative (Fig. 5). Metastatic neoplastic cells in the sublumbar lymph node had similar cytoplasmic immunoreactivity.
Anal sac gland carcinoma; cat. Immunohistochemistry for CAM 5.2. Closely packed immunolabeled cells line tubuloacinar structures and form serpiginous trabeculae. Bar = 100 µm.
Anal sac gland carcinoma; cat. Immunohistochemistry for CAM 5.2. Acini and cords of neoplastic epithelial cells within the tumor (T) are strongly immunolabeled. Hyperplastic and dilated apocrine sweat glands within the adjacent anal sac wall exhibit similar immunoreactivity (asterisks). Bar = 0.3 mm.
Anal sac gland carcinoma; cat. Immunohistochemistry for CAM 5.2. A deep nodular neoplasm is outlined by immunolabeling (T). Epidermis and sebaceous glands (asterisks) do not react with this marker, whereas cutaneous apocrine sweat glands exhibit strong immunoreactivity (arrows). Bar = 0.5 mm.
Three canine formalin-fixed paraffin-embedded tissues of hepatoid gland (perianal gland) origin (normal perianal glands, benign perianal adenoma, and perianal carcinoma) did not express CAM 5.2 (unpublished data).
Clinical outcome
Of the 64 cats, postsurgical outcome was known for 39 cats. Twenty-nine cats initially underwent surgical resection or debulking; 17 cats had only an incisional biopsy performed; surgical intervention was unclassified in the remaining 18 cats.
Five cats were euthanatized within the perioperative period because of surgical complications, such as hemorrhage/hematoma formation, poor wound healing, or constipation. Euthanasia was elected by the owner in 2 cases after receipt of the histologic diagnosis and accompanying unfavorable prognosis.
Tumor progression (n = 6) or localized recurrence (n = 16) were common complications, eventually necessitating euthanasia, mainly because of local infiltrative growth and impairment of defecation. The time for tumor regrowth varied from 1 to 16 months.
Presumptive metastatic lesions in liver, lung, abdominal lymph nodes, and diaphragm were detected by physical examination, radiographically, or both in 6 cats, but metastatic disease was not confirmed postmortem in any of these cats. In 1 cat, sublumbar lymph node metastasis was documented histologically at the time of initial tumor resection.
Three cats died or were euthanatized because of unrelated causes; 3 cats diagnosed during 2005 and 2006 had no evidence of tumor recurrence or metastasis at the most recent follow-up and were presumably still alive at the time of this writing. For the 33 of 39 (84.6%) cats that were known to have been euthanatized or died because of clinical problems directly attributable to the primary tumor or presumptive metastatic disease, 1- and 2-year survival rates were 19% and 0%, respectively. Median survival was 3 months (range, 0–23 months).
Discussion
Anal sac gland carcinoma is an uncommon and aggressive malignant tumor that arises from the apocrine glandular tissue of the anal sacs. It has been well characterized in dogs but rarely reported in cats. To the authors' knowledge, only 2 cases have been published previously.
One case involved a 12-year-old neutered male Siamese cat with a history of constipation and surgical resection of an enlarged anal sac. Recurrence in the form of perianal cutaneous nodules was noted 4 months after surgery, and the cat was euthanatized at 6 months because of ongoing constipation. 15 The other case was in a 15-year-old neutered female DSH with a history of a painful perianal swelling. Surgical resection was performed, but long-term follow-up was not available for this case. 18
This report describes ASGC in 64 cats, diagnosed over a 13-year period, from surgical biopsy submissions received by a private diagnostic laboratory in the UK. In addition, the report provides information on the clinical outcome in 39 cats. The geographic distribution of feline ASGC did not differ significantly from that of the reference population.
In the dog, ASGC has been reported mainly in older females. (The average age at presentation is approximately 10 years. 9,19 ) However, recent studies indicate a roughly equal gender distribution. 21,25 The mean age at tumor diagnosis of cats in this study was 12 years, with a range of 6–17 years. This is consistent with this being a neoplasm of mature to aged individuals. Sixty-one percent of the cats with ASGC were female, in contrast to the approximately equal gender distribution for all feline skin neoplasms and total feline biopsy accessions. This could suggest female predilection; however, any inferences regarding gender or sex hormonal influence are tempered by the fact that most cats in this study were neutered.
The DSH was the most frequently affected breed (81.1%); Siamese cats (7.8%) were the next most commonly affected breed; few cases were observed in other breeds. This distribution could reflect the relative popularity of these cat breeds, rather than indicating any genetic or familial predisposition. However, the proportion of Siamese cats among cats with ASGC was 4 times the proportion of Siamese cats among those with all skin neoplasms and almost 3 times that among the laboratory reference population. In dogs, several authors have suggested increased breed-associated risk for ASGC, especially in the Cocker Spaniel. 8,20,21
In dogs, ASGC generally presents as a poorly defined swelling or mass arising from the anal sacs. Tumors often invade deeply into perirectal tissue, but ulceration is uncommon. 8 The cats in our study had similar clinical presentation; however, ulceration and the development of fistulae were frequent complications in feline ASGC. A number of tumors were initially treated by the attending veterinarian as anal sacculitis/anal sac abscess and only subsequently investigated and diagnosed following poor response to conservative therapy. Unlike dogs, in which anal sac impaction, subsequent infection and abscessation is relatively common, cats seldom present with anal sac disease, and impaction generally occurs without infection. 23 Thus, ASGC should be included in the differential diagnosis for any older cat presenting with clinical signs of anal sac disease.
In dogs, invasive behavior of ASGC is frequent. 7,12,24 These tumors commonly metastasize to regional lymph nodes and less frequently to other abdominal organs and the lungs. 3,16,22,25 In this study, most feline ASGC infiltrated capsular fibrous tissue and perirectal soft tissues and occasionally invaded the rectal wall or peritumoral lymphatics. Tumors were often debulked, rather than completely resected, and the recurrence rate was high. Presumptive metastatic disease, involving sublumbar lymph nodes and abdominal, thoracic, or both viscera, was documented in approximately 16% of cats; however, in only 1 cat was regional lymph node metastasis confirmed histologically.
In a study of 43 dogs with ASGC, 79% had evidence of metastatic disease at the time of clinical presentation, with the regional lymph nodes (72%) being the most common site of metastasis. 3 These data suggest that ASGC might behave more as a locally aggressive tumor in the cat, with less propensity for metastatic dissemination. However, this disparity might simply reflect differences in the degree of clinical investigation and tumor staging.
Sciatic nerve paralysis was detected in 1 cat with recurrent tumor growth. The exact cause was undetermined, but tumor invasion cannot be ruled out because metastatic anal sac adenocarcinoma has been reported to cause hind limb paresis that rapidly progressed to paralysis in a dog. 4
Median postsurgical survival time for cats in this study was 3 months, with only 19% of cats surviving beyond 1 year and none surviving to 2 years. These figures exclude the cats for which surgical resection was apparently curative or that died or were euthanatized for unrelated causes. These data are similar to those in a canine study, in which the median survival for a group of 43 dogs was 6 months. 3 However, direct comparison might not be appropriate because most of the dogs were treated for metastatic disease and hypercalcemia, whereas most cats in this study were euthanatized because of local recurrence or postsurgical complications. A larger canine study (n = 113) examining various treatment modalities (surgery, radiation, chemotherapy) revealed a median survival of 544 days for treated dogs, suggesting a more favorable prognosis than previously reported. 25 In this study, cats were treated surgically; only the 1 cat with confirmed sublumbar lymph node metastasis received chemotherapy.
Hypercalcemia of malignancy (pseudohyperparathyroidism) associated with tumor secretion of PTHrP has been variably reported (25–90%) in association with canine ASGC, 3,16,23 although there may be some discrepancy between the number of hypercalcemic dogs and those that actually exhibit clinical signs, such as polyuria-polydypsia. 3 In this study, although a number of cats had preanesthetic serum biochemistry profiles, serum calcium was rarely included in the panel. Overall, serum calcium concentration was available for 5 cases; only 1 cat had (marginal) hypercalcemia. No clinical signs attributable to hypercalcemia were observed in this cat, and no further follow-up investigation was performed. Interspecies variations might exist in the ability of anal sac carcinomas to synthesize and secrete humoral hypercalcemic substances. For example, in 60 ASGC in ranch mink, the tumors were not accompanied by hypercalcemia. 11 Moreover, serum calcium concentration was within reference limits in the aforementioned case in a Siamese cat. 15 However, the limited number of cats for which serum calcium concentration was measured precludes any definitive conclusions regarding the occurrence of humoral hypercalcemia in feline ASGC.
As a part of this study, we investigated the utility of an immunohistochemical marker (CAM 5.2) for the diagnosis of ASGC in a randomly selected subset of tumors. CAM 5.2 is a murine monoclonal antibody against the colon carcinoma cell line HT29. This marker recognizes lower molecular weight intracellular cytokeratin proteins within secretory epithelia (primarily CK8 and CK7) but does not react with stratified squamous epithelium. 10,19 This marker has been employed mainly in human medicine for classifying carcinomas and for distinguishing poorly differentiated carcinomas from malignant tumors of nonepithelial origin. 1,17 CAM 5.2 has been used in veterinary medicine in the diagnosis of cutaneous apocrine sweat gland tumors. 10 In all 12 cases tested, the neoplastic cells of feline ASGC were immunohistochemically positive for cytokeratin CAM 5.2. Stratified squamous epithelium of the anal sac lining and epidermis, hair follicles, and sebaceous glands, which were included in some histologic sections, did not express this marker. Moreover, antibody to CAM 5.2 did not react with normal or neoplastic canine hepatoid gland (perianal gland) tissue (unpublished data). Considering these results, this antibody could be useful for the immunohistochemical classification of poorly differentiated tumors of the perianal region in both dogs and cats.
Although this study has several limitations, many of which were unavoidable given the retrospective nature of the investigation and the multiple primary care veterinary practices involved (e.g., limited clinical investigation, lack of archived case records and necropsy findings), it has yielded useful information about the clinical manifestations and biological behavior of ASGC in the cat and raises awareness of this condition. On the basis of the cases in this study, a guarded to poor prognosis seems appropriate, depending on early detection and completeness of excision. This report provides evidence that, although rare, ASGC should be recognized as a distinct clinical entity in the cat and should be included in the differential diagnosis for older cats with clinical signs of anal sac disease.
Footnotes
Acknowledgements
We thank all practitioners involved in this study for their collaboration, without which this study would not have been possible. In addition, we thank Mr. Martyn Carpenter of Finn Pathologists, Weybread, Diss, UK, for financial support and offer special thanks to Ian Archer for performing the immunohistochemistry and preparation of slides and Dr. David Hardy of Pfizer Ltd., Sandwich, Kent, UK, for performing statistical analysis.